Erlang

Erlang Solutions: Handling spikes & loads in a new era of gaming

The bookmakers of the future may not be who you’d expect, writes Stuart Whitfield, CEO at Erlang Solutions.

Stuart Whitfield, Erlang Solutions

Chris Eaton, industry expert, integrity monitor and former investigator for INTERPOL is quoted as saying: In 10 years’ time, I expect sports betting to be part of major telecommunications companies and data companies.

Technology and telecommunications companies deal with massive volumes of data and transactions in real-time, 24/7. As online gambling and betting itself moves into a realm of ever-increasing real-time, data-led transactions, success depends on being able to harness and deploy the appropriate technology and market data.

In short, it’s time for those operating (or aspiring to operate) in the sports betting space to think and act like a data / telecoms / technology company. They can achieve this a number of ways, but the cornerstones of this approach are likely to share some common characteristics.

To paraphrase G. K. Adams, to go forwards, we must first look back. Where has online gambling and betting come from, and what changes need to be made to go where they’re going?

Here are some key considerations for those wishing to succeed in this burgeoning market:

Traditional sports to esports

With World Cup 2018 just around the corner, it seems unthinkable that football (or soccer, for the subjects of Trump) could be unseated as the world’s biggest betting draw. But that’s a very real possibility with the exponential growth of esports, which will both compliment and disrupt the traditional sports betting industry.

Today’s kids who take part in and follow competitive gaming will be the punters of the future. This new breed of punter may like football, but they may play and wager on it through esports, together with a range of increasingly popular esports titles (Dota2, League of Legends, Counter Strike: Global Offensive, Street Fighter etc) . This shift will have a huge impact on the sports industry in the long term.

And it already is; major esports events were broadcast on mainstream American cable networks for the first time in 2016. By 2018, esports is expected to be a $1B business1. With this, peak traffic will come more often, on a global scale. In this new phase of the industry real-time scalability, concurrency, and rock-solid reliability can make or break a betting operator, as  in finance and telecoms.

Monolith to distributed microservices

Betting systems have often followed the monolith model, where all services had a central governance, typically employed on a single platform and with many dependencies. Scaling up for peak traffic was an all-or-nothing effort. A canonical data model was normally used. Today, when considering scalability, maintenance, and innovation, this approach is not considered to be very effective.

Monoliths are very hard to extend, mainly because of the huge codebase and long deployment process. In addition to this, because of the massive inertia of many big betting systems, it’s very hard to extend the monolithic system with technologies that might be better at serving certain features.

The basic idea of a micro-services architecture is to have many small services that do one (and only one) thing well. Applications are put together by defining a group of distinct services and how they interact. For any given service we can vary how many instances of that service is running.

Distributed microservices allow instances of services to be dynamically distributed across different servers for redundancy and/or to scale up/down to fit the capacity of the service. Instances can be moved between servers or even among competing server providers to reduce runtime costs. Hence scaling up or down at peak times isn’t just about computational capacity, but also includes savings in running costs.

SQL to NoSQL

There were several competing databases (network dbase, graph dbase, even flat file) but relational databases were hugely popular mainly due to SQL. SQL gave a structured means of creating, inserting and querying the relational database. SQL became the lingua franca of data modelling and was embraced by many established betting shops.

The initial NoSQL servers were simple caches of previously computed values to reduce redoing fix calculations. These caches were quickly expanded over a network of servers to account for the needs of web applications. It was found this was simpler and cheaper to achieve compared to relational SQL databases. As more data was moved into these caches the first formal NoSQL databases took form.

The best tool for the job

The software development community is going through a revolution of diversity and inclusion as more developers move away from the biggest programming languages in search of the right tool for the nuanced challenges of our new, technologically-literate world. The result: it’s becoming rarer for there to be de facto programming languages that organisations are forced to use because that’s what their developers community uses.

One-size-fits-all stacks will never scale indefinitely. The most important lesson we’ve learned working in the gaming industry has been realising that a silver bullet for solving every problem simply does not exist, and it’s important to pick a proper tool for the job. Erlang and Elixir are superb at scaling reliably and efficiently when betting operators need it to, but also complement other tools and languages which are better suited to other aspects of a system.

One needs to look no further than bet365, whose Systems Development Manager for Middleware, Andrew Deane, recently went on the record to say “When I joined bet365 4 years ago, the company was in its second wave of Erlang development.

Having proved the model with Erlang, the door was open to explore other languages. Continuing the mantra of the right tool for the right job, we’ve found use cases for Google’s Go in our notifications systems and testing tools and more recently for Elixir in our application tooling.”

We are trusted partners to some of the world’s leading and most innovative online gambling and betting operators. Learn more about Erlang Solutions and download our eBook: A Guide to Handling Massive Spikes and Loads.